Ma, Chenchen, authorFassnacht, Steven, advisorKampf, Stephanie, advisorWei, Yu, committee member2017-09-142017-09-142017https://hdl.handle.net/10217/183907In many high elevation mountain regions, documented warming rates have been greater than the global surface average. These warming rates directly affect the snowpack, runoff, ecosystems, agriculture and species that rely on a high elevation snowpack. Temperature records from the snow telemetry (SNOTEL) network across the Southern Rocky Mountains in the western United States have high warming rates, which may have been affected by systematic inhomogeneities in the temperature data caused by sensor changes. This study evaluates the maximum, average, and minimum temperature trends from 68 long-term SNOTEL stations across Colorado for the period from the 1980s through 2015 using the non-parametric Mann-Kendall/Theil-Sen's analyses before and after the temperature records were corrected for the sensor-caused inhomogeneities. Three homogenization methods were tested using a simple temperature index snow accumulation and melt model. Results show that the significant warming trends found in the original datasets, especially in minimum temperature (average increase of 1.2 °C per decade), decreased (to an average of 0.5 °C per decade) after homogenization. Step-like shifts in temperature datasets were observed in SNOTEL temperature records at the time of temperature sensor change, which created a discontinuity in the temperature dataset. The temperature-index snow model simulated snow water equivalent (SWE) well (more than 93% of the calibrated stations within the "good" and "very good" performance category for all three statistical-evaluation periods based on the Nash-Sutcliffe coefficient of efficiency, NSCE) using the new temperature sensor dataset. However, these models did not perform as well when using the original (pre-sensor change) and homogenized temperatures, with 23% of stations for the original temperature data and 44-69% of stations for two homogenized temperature datasets within the "good" and "very good"temperature data, but they did not fully correct for the effects of sensor change on the temperature records. The NSCE and bias statistics from SWE modeling using the original and homogenized datasets suggest that the homogenization methods evaluated in this study are applicable for many of the SNOTEL stations in Colorado but not all, and need to be applied with caution. Potential users of temperature products from the SNOTEL network should also be very careful when choosing time periods for future climate change research and assessments. More long-term climate monitoring stations should be installed in high elevation mountain regions to document and investigate elevation-dependent warming.born digitalmasters thesesengCopyright and other restrictions may apply. User is responsible for compliance with all applicable laws. For information about copyright law, please see https://libguides.colostate.edu/copyright.Text